رفتار سازه آلومینیوم اتصالات پرچ شده خود نافذ: یک تحقیق تجربی و عددی

The present paper deals with the structural behaviour of self-piercing riveted joints based on aluminium and steel rivets. Two T-components made of two open aluminium profiles in alloy AA6063 temper T4 joined by 6 and 12 rivets, respectively, were designed and tested under quasi-static loading conditions. A new test device was designed to perform the tests of the T-components under two different load cases. Experimental results of the T-components joined by using aluminium self-piercing rivets were then compared with the corresponding components joined by using steel rivets in terms of force-displacement curves, deformation modes of the components as well as rivet failure modes. Further, the experimental results of the T-components based on aluminium rivets were used to validate a resultant-based point-connector model for self-piercing rivets proposed by Hanssen et al. (2010) using shell elements.

The self-piercing riveting (SPR) technique is an alternative to the welding technology and is nowadays widely used in the automotive industry. Significant knowledge about the behaviour of a SPR connection under static and dynamic loading conditions (including fatigue) can be found in the open literature (Fu and Mallick, 2003, Han et al., 2007, Han et al., 2010, Hoang et al., 2011, Hoang et al., 2010, Lee et al., 2006, Li and Fatemi, 2006, Mori et al., 2006, Porcaro et al., 2004, Porcaro et al., 2006, Porcaro et al., 2008, Sun and Khaleel, 2007, Sun et al., 2007 and Wood et al., 2011). Recently, self-piercing riveted connections based on a single aluminium rivet have been studied by Hoang et al. (2010) and Abe et al. (2009) in order to facilitate recycling of an aluminium car body in the future by reducing the unfavourable iron content. Be aware that unfavourable alloy content could also be the case if we do not adjust the alloy of the aluminium rivet to the aluminium alloy to be joined. Moreover, the possibility of using aluminium rivets as an alternative to steel rivets can contribute to the reduction of the car weight. A quick calculation reveals that the substitution of steel rivets with aluminium ones in the body of a Jaguar XJ, in which more than three thousands rivets are presents, can save approximately 1 kg of weight. Finally, the use of aluminium rivets to join aluminium plates solve also problems related to galvanic corrosion which is imminent when using steel rivets to join aluminium sheets as stated by He et al. (2008). The work of Hoang et al. (2010) have shown that the behaviour of a connection using an aluminium rivet to connect two aluminium sheets was similar to that of using steel rivets in terms of initial stiffness, maximum strength, and softening behaviour after the onset of failure. Their findings showed a great potential of aluminium self-piercing rivets for replacing the steel ones. However, in order to push forward the application of aluminium rivets in the automotive industry, research has to be carried out in order to better understand the structural behaviour of aluminium riveted joints. In addition, a reliable point-connector model is needed to describe the local behaviour of the riveted connection in a reasonable way for full car crash simulations with shell-element based models. Traditional approaches (e.g., node-to-node constraints, node-to-surface and surface-to-surface constraining by contact formulations, using beam elements, brick elements for the connector, etc.) were basically developed for spot welded connections, and may be used for modelling self-piercing riveted connections. However, the physical mechanisms during the failure of a rivet connection are complex, and completely different from that of a welded one. Thus, the application of traditional approaches for riveted connections may not give satisfactory results. In this context the work of Porcaro et al. (2004) can be mentioned. They used a node-to-node constraint approach for modelling single self-piercing riveted connectors, and obtained reasonable results up to maximum load. However, the softening behaviour of the riveted connection beyond maximum load was not correctly described, neither with a force-based failure criterion nor with a strain-based failure criterion. Recently, (Hanssen et al., 2010) have developed a new resultant-based point connector model for large-scale finite element shell analyses. The nature of the model is based on the observed physical failure mechanisms of a self-piercing rivet connecting two aluminium sheets (Hanssen et al., 2010). They showed that the model was able to capture with good accuracy the behaviour of riveted connections with a single rivet up to failure for different loading directions, for different aluminium sheet thicknesses as well as for different rivets and die types. However, the capability of their proposed model for modelling the structural behaviour of self-piercing riveted joints remains an open question. Thus, in the present study the structural behaviour of aluminium self-piercing riveted joints by using T-component tests was first investigated experimentally. T-components have been commonly used to investigate the joint behaviour in many research works found in the open literature Clarke et al., 2009, Díaz et al., 2011, Girão Coelho and Bijlaard, 2007, Jones et al., 1983, Seeger et al., 2008, Swanson et al., 2002 and Vivio, 2009. However, most of them were designed to investigate structural behaviour of bolted joints, welded joints, and adhesively bonded joints. Within this study, two new T-component specimens were designed, adapting to the SPR process. The geometry was chosen as a function of the joining accessibility, loading complexity, and expected structural behaviour. They consisted of two open profiles in aluminium alloy AA6063 in temper T4 which were joined together by using 6 and 12 aluminium self-piercing rivets in alloy AA7278-T6, respectively. Here, a commercially available high strength aluminium alloy was chosen to facilitate the riveting process, and to study the structural behaviour of aluminium joints without any coupling to recycling. Two load cases were investigated in order to challenge the riveted connectors to different failure modes. The experimental results using aluminium rivets were compared with the test data from the same T-component, but joined by using steel self-piercing rivets. The test data using aluminium rivets were finally used to validate the point-connector model proposed by Hanssen et al. (2010).

The main objective of the present paper was to investigate the structural behaviour of the T-components made by joining two aluminium extrusions using aluminium self-piercing rivets, and to check the capability of the SPR model proposed by Hanssen et al. (2010) for large scale shell analyses. The study was carried out experimentally and numerically. The main findings of the present work can be summarised as follows. • The overall structural behaviour of the T-components by using aluminium rivets under different loading conditions was comparable to those by using steel rivets. This included also the failure mode of the rivets. • The components using aluminium rivets showed a maximum force level which was approximately 5–8% lower than for the components using steel rivets. • The ductility of the T-components based on aluminium rivets was approximately 50% less than those based on steel rivets. The lower ductility resulted in a lower energy absorption capacity of the self-piercing riveted joints based on aluminium rivets. • Numerical analyses revealed that the point-connector SPR model was able to predict quite well the global behaviour of the T-components when using aluminium self-piercing rivets in terms of force-displacement curves, deformation of the profiles, as well as rivet failure. However, for complex loading conditions of the rivets (large local deformations of the profiles coupled with shear and tension loading of the rivets) the SPR model was not able to give a correct prediction of the structural behaviour of the riveted joints (e.g. force-displacement curves as well as the failure of individual rivets, etc.). The reason for this is not quite understood, but could be caused by the anisotropy of the aluminium profiles which is not taken into account in the calibration of the SPR model. • The numerical analyses showed that the peeling tests used to calibrate the SPR model parameters had minor influence on the structural behaviour of the T-components. Thus, it seems the structural behaviour of self-piercing riveted joints can be predicted with reasonable accuracy by using only three U-shaped specimen tests for calibration. • The effect of anisotropy of the profile material in the simulations (i.e. by using an isotropic and anisotropic yield criterion) seemed to have minor effect on the structural response. However, the anisotropy of the profile material may result to anisotropy of the riveted connection strength with respect to the loading direction (see Fig. 13), which needs to be taken into account in the SPR model in future work.